作者: BRUNERO LISEO
关键词: Prior probability 、 Statistics 、 Mathematics 、 Bayesian probability 、 Frequentist inference 、 Econometrics 、 Nuisance
摘要: SUMMARY The problem of eliminating nuisance parameters is tackled from different points view. Standard likelihood techniques, such as profile and its modifications, are compared with a Bayesian analysis based on reference priors (Berger & Bernardo, 1992a). Examples considered to illustrate what happens in multiparameter problems. In particular the frequentist coverage probabilities credible sets derived shown be better than those computed approach. Advantages disadvantages analyses discussed.